The various injection modes in the Sample Manager act in different ways, and use different mechanisms, which means they will give different responses, although within one mode the effect will be reproducible, the areas will be different. Here, you are seeing a difference in recovery, which is likely related to a mismatch in the method settings. If you see a peak with full loop then clearly the injector is working, (I am assuming the peak is reproducible), but as the peak is small in partial loop mode, then the method needs modification. I cannot tell if you are using PLNO or PLPA mode. Can you send details on the whole method, sample diluent and needlewashes so that I can make a recommendation.
Also there are some additional guidance on this on some of the threads already on the community, for you to review. I will conclude this post with an attachment explaining factors that affect the accuracy more detail.